Getting plate-solving status, please wait...

Ghosts in Cepheus

Technical card

Resolution: 2863x2045

Dates: Oct. 17, 2015

Frames: 30x420" ISO1600

Integration: 3.5 hours

Darks: ~30

Flats: ~30

Bias: ~30

Avg. Moon age: 3.64 days

Avg. Moon phase: 14.27% job: 911219

RA center: 319.103 degrees

DEC center: 68.234 degrees

Pixel scale: 0.930 arcsec/pixel

Orientation: 2.874 degrees

Field radius: 0.455 degrees

Locations: Staunton River State Park, Red Oak, VA, United States


Here is a shot at the Ghost Nebula in Cepheus, captured at the Staunton River Star Party last October. I completely underestimated the faintness of the dust in this region! Originally, I had planned to dedicate two nights to this object and acquire piles of integration time, but the skies were dark and the temperatures were damn cold so I figured 3.5 hours was enough. Nope! Not even close! The dust is so faint that I couldn’t push it too hard to separate it from the background without the image becoming a tattered noisy mess. Fortunately, I used a multi-scale processing technique by Emanuele Todini which helped enhance the contrast between the dust and background with only moderate increases in noise (step 13 below). I did notice that the halos around the brightest stars increased during this process, but I did nothing to mitigate the effect. Regardless, I am very excited to experiment more with multi-scale processing which is what PixInsight is built for.

All told, I am disappointed with the SNR of this one. It really stands out because this is a pretty tight crop of my larger FOV. Later this year I plan to dedicate more time to this object hopefully in time for Halloween.

Original - cropped final image
Revision B - inverted monochrome
Revision C - closeup of the ghosts
Revision D - uncropped image

Processing Workflow (PixInsight)

1. Initial crop (Dynamic crop) .
2. Neutralize background (BackgroundNeutralization).
3. Set white balance (ColorCalibration).
4. Set luminance coefficients to 0.333333 for RGB channels (RGBWorkingSpace).
5. Deconvolution with luminance mask.
6. Reduce background noise (TGVDenoise with inverted luminance mask).
7. Reduce background luminance noise (MultiscaleMedianTransform to Luminance)
8. Reduce background chrominance noise (MultiscaleMedianTransform to Chrominance)
9. Non-linear stretch (HistogramTransformation, lower midtones slider aggressively).
10. Lower blackpoint (HistogramTransformation, raise blackpoint slider).
11. Reduce bright pixel noise with Manfred Schwarz’s script (RemoveHotPixel)
12. Darken image slightly (CurvesTransformation to RGB/k)
13. Increase brightness, color saturation and contrast of dust using Emanuele Todini’s multiscale processing technique (reference below)
13a. Duplicate image and extract small scale structures (ATrousWaveletTransform, 5 layers, deselect residual)
13b. Create new image of large scale structures by subtracting previous image from original image (PixelMath, deselect rescale).
13c. Stretch image of large scale structures (HistogramTransformation, lower midtone slider with white and black clipped luminance mask from original image selecting dust)
13d. Increase color saturation of large scale structures (CurvesTransformation to saturation with previous mask).
13e. Merge images of large scale structures, small scale structures and original image together (PixelMath, rescale selected).
13f. Repeat the procedures above except using white and black clipped luminance mask from large scale structure image image selecting dust and additionally adding contrast curve.
14. Reduce green (SCNR to green, amount: 0.8. interestingly, amount = 0.63 left a slight but sickly sludge green cast - that was the original version - and amount = 1.0 compromised the yellow of the dust too much. Amount = 0.8 seemed to allow the dust to look warm without the green cast).
15. Increase color saturation (CurvesTransformation to Saturation with luminance mask).
16. Reduce star sizes (MorphologicalTransformation with range mask - MorphologicalSelection iterations 3, amount 0.6, selection 0.18, 7 element diamond structure).
17. Reduce dark pixel noise with Manfred Schwarz’s script (RemoveDarkPixel)
18. Sharpen dust (MultscaleMedianTransform to Luminance, add bias to layers 4 (+0.05) and 5 (+0.05) with luminance mask selecting nebula only).
19. Final histogram tweaks to balance RGB of background (HistogramTransformation).
20. Reduce background brightness to median value of 0.1 (CurvesTransformation and blackpoint in HistogramTransformation).
21. Final crop to 5 x 7 aspect ratio (DynamicCrop)
22. Set ICC profile to sRGB for web publishing (ICCProfileTransformation).


“M42 PixInsight Tutorial” by David Ault for linear noise reduction

“M42 & NGC1975 – Multilayer Processing (First Part)” Tutorial by Emanuele Todini

Manfred Schwarz’ galaxy tutorial:



Jason Tackett


Sky plot

Sky plot


Ghosts in Cepheus, Jason Tackett

Made possible by

O'Telescope BackyardEOS